Re: Partitioning for Speed - heads
hi ya oki
On Tue, 1 Apr 2003, Oki DZ wrote:
> On Mon, Mar 31, 2003 at 05:03:17PM -0800, Alvin Oga wrote:
> > - less moving of the heads is generally faster as it takes too long
> > to move the head relative to just reading data
>
> How long is exactly to move the head from outside to the center?
> Long ago, I tinkered with moving the swap partition inside/outside, but
> now I believe that the difference is hardly noticeable.
nobody shoould be moving heads from the outside ( track 0 ) all the way
to the inside... that'd be ridiculously slow
- average seek latency is the magic number what you;re looking for
-- but thats just average .. who knows what that means...
- say its 3ms or as high as 100ms....
and the disks is spinning round-n-round at 7200rpm or 10000rpm
or 15000rpm would be a lot of missed data while moving heads
that you could have read on all 8-16 tracks (number of real heads)
- a good ide driver will improve disk i/o ... :-)
that maybe we can finally get the 100MB/sec out of a ata-100 disk
- raid comes closer/better with disk reads .. much faster than a single
disk
- notice ...
- as the disks get full ... performance degrades
- as you add more disks ... your performance increases
http://www6.tomshardware.com/storage/20000329/fastrak66-14.html
- zone bit recording
http://www.pcguide.com/ref/hdd/geom/tracks_ZBR.htm
c ya
alvin
Reply to: